Deep Vision and Graphics

This repo supplements course "Deep Vision and Graphics" taught at YSDA @fall'26. The course is the successor of "Deep Learning" course taught at YSDA in 2015-2021. New course focuses more on applications of deep learning for computer vision.

Lecture and seminar materials for each week are in ./week* folders. Homeworks are in ./homework* folders.

General info

• YSDA deadlines & admin stuff can be found at the YSDA LMS (ysda students only).
• Any technical issues, ideas, bugs in course materials, contribution ideas - add an issue

Syllabus

• week01 Intro, recap of Neural network basics, optimization, backprop, biological networks, images, linear filtering, convolutional networks, batchnorms, augmentations
• week02 ConvNet architectures and how to find them, sparse convolutions in 3D, ConvNets for videos, transfer learning
• week03 Non-convolutional architectures: transformers (some recap of their use in NLP), mixers, FFT convolutions
• week04 Semantic segmentation
• week05 Object detection, instance/panoptic segmentation
• week06 Representation learning: face recognition, verification tasks, self-supervised learning, zero-shot learning
• week07 Latent models (GLO, AEs, VQ-VAE). Generative adversarial networks
• week08 Diffusion models, generative transformers
• week09 Multimodal networks
• week10 Pointcloud processing networks
• week11 Shape and motion estimation: spatial transformers, optical flow, stereo, monodepth, point cloud generation, implicit and semi-implicit shape representations
• week12 New view synthesis: multi-plane images, neural radiance fields, mesh-based and point-based representations for NVS, neural renderers

Contributors & course staff

Course materials and teaching performed by

• Victor Yurchenko - lectures, seminars, homeworks, admin stuff
• Fedor Ratnikov - lectures, seminars, homeworks, admin staff
• Kirill Struminsky - lectures 11&12, seminars, homeworks
• Alexander Timofeev-Karakozov - seminars on detection and segmentation
• Egor Kolesnikov - seminar on vision transformers
• To be continued